Characterizing an arm using a neo550 and a through bore

Bmongar · February 20, 2023, 12:40pm

Today we will be characterizing the wrist of our intake it has a neo550 at a XX to YY gear ratio, we have a through bore encoder attached to the shaft of the wrist that is attached to the DIO on the Roborio. (that is how use use a REv through bore for absolute right? Dudycycle encoder on dio?)

Trying to figure out what the settings in sysid need to be. I know we need to limit sysid because the wrist will move fast. we were able to see movement at .12% output

I am not sure what the encoder selection should be, I don’t see a place to use the dio. or a place to oput the port number on any of the settings.

Then for units per rotation, in the logger that is the motor or the encoder?

Joe_Ross · February 20, 2023, 4:07pm

SysID doesn’t appear to have a setting for a duty cycle encoder

That was the only way prior to 2023. However, I don’t think SysID supports the absolute encoder mode.

calcmogul · February 20, 2023, 5:11pm

Correct. SysId needs velocity measurements that aren’t provided natively by duty cycle encoders. I’ve been experimenting with a different OLS formulation for SysId that supports position data as well:

github.com

calcmogul/SysIdOpt/blob/main/src/Main.cpp#L115-L266


      
          /**
           * Solves linear system ID problem with Eigen to produce initial guess for
           * nonlinear problem.
           *
           * @param[in] json SysId JSON.
           * @param[in] motionThreshold Data with velocities closer to zero than this are
           *   ignored.
           * @param[in] positionStddev Position standard deviation.
           * @param[in] velocityStddev Velocity standard deviation.
           * @return Initial guess for nonlinear problem.
           */
          FeedforwardGains SolveEigenLinearSystem(
              const wpi::json& json, units::meters_per_second_t motionThreshold,
              units::meter_t positionStddev, units::meters_per_second_t velocityStddev) {
            // [p]    = [1  a][p]  + [0]   + [0]
            // [v]ₖ₊₁   [0  b][v]ₖ   [c]uₖ   [d]sgn(vₖ)
            //
            // [p]    = [pₖ] + [vₖa] + [ 0 ] + [   0    ]
            // [v]ₖ₊₁   [0 ]   [vₖb]   [uₖc]   [sgn(vₖ)d]
            //

This file has been truncated. show original

Kartikr27 · February 22, 2023, 7:01pm

Since Sysid does not support the absolute encoder mode, is it best to just use the integrated encoder on the NEO to use characterization?

Bmongar · February 24, 2023, 1:13am

I am trying to figure out the characterization settings for an arm, i decided to use the falcon encoder for the purposes of characterization. I can’t change the 6.2832 on the units per rotation and don’t know where it came from.

Also trying to determine a good setting for dynamic step voltage

Joe_Ross · February 24, 2023, 2:34am

That’s the number of radians in 1 rotation (2pi)

Bmongar · February 24, 2023, 9:14am

Duh! Looking at the bot too long. Tau did I ever miss that?

sk2983 · February 24, 2023, 12:53pm

you can run absolute through the spark max with either the absolute encoder breakout board or a custom wiring harness.